The following paper by David Packham was provided as Submission No 10 for the Inquiry into the Wambelong Fire on 28/01/2014.
We have republished this paper without the reference notes to better suit web site viewing. Readers are encouraged to download the original paper that includes reference notes. Click HERE to download that document in a pdf format.
Some observations on the effectiveness of fuel reduction burning in Southern Australia
By David Packham – 16 August 2010
Introduction
The recent Royal Commission into the fires of 7 Feb 2009 which in terms of life loss was the worst natural disaster in Australia recommended a fuel reduction program that amounted to 5% of publically owned land in Victoria. A crucial question is what can be expected in terms of decreased loss of life and damage as a result of such a fuel management strategy? What would be the effect of prescribed burning levels other than 5%, perhaps 10% and 15%? This note is an attempt to provide some guidance to this crucial question.
Methodology
An examination of life loss in disaster fires in southern Australia and modelled and actual relationships between prescribed burning and subsequent bushfire area burnt has offered the opportunity to estimate the efficacy of prescribed burning as a protective strategy.
It is assumed that a decrease in fire intensity and fire area will result in a decrease in casualties and asset damage. A correlation between average Byram Fire Intensity (mega watts per meter) will be sought for disaster and extreme fires that involve both rural and urban environments. The average fuel levels for different fuel reduction regimes will be estimated and the life loss – fire intensity equation used to compare expected life loss for extreme or catastrophic fire for 5, 10 and 15% annual prescription burning.
A different strategy will be the use of actual and modelled annual average area burnt by wildfire as a function of fuel reduction.
Life loss as function of Byram fire intensity index
In this analysis the forecast or recorded fire weather conditions and estimates of fuel have been correlated with the loss of life for notable disaster fires that have impinged upon populated areas. That correlation produces a simple linear equation that allows prediction of life loss for the same fire weather but reduced fuel as a result of chosen fuel reduction fire frequency.
Table 1. Gives a FFDI for six disaster fires in South East Australia with the concurrent loss of life. A simple linear equation is fitted.
| Fire Event | Peak FFDI | Estimated Byram Fire Intensity Index, Megawatts per metre | Lives lost |
| Black Friday 1939 | 100 | 29 | 71 |
| Victoria 1944 | 110 | 41 | 49 |
| Hobart 1967 | 78 | 42 | 62 |
| Ash Wednesday Victoria 1983 | 191 | 72 | 73 |
| ACT 18 Jan 2003 | 105 | 47 | 4 |
| Black Saturday, Victoria, 2009 | 164 | 89 | 173 |
The selected fires give a relationship between the Fire Intensity and Lives lost as
Lives lost = 0.66 x Fire intensity (mega watts per metre)– 11 r=.52
The simple equation has been applied to bushfires but burning in landscapes that have had different fuel reduction burning frequencies. The nominal Forest Fire Danger Index (FFDI) of 100 is chosen as it represents the upper limit of extreme fire weather and an index that can be expected a few times per decade.
Table 2. Gives the estimated fine fuel levels and estimated life loss for a FFDI of 100 with five fuel reduction regimes.
| Fuel reduction burning regime percent per annum | Estimated average fine fuel levels, tonnes per hectare | Estimated Byram Fire Intensity Index for a FFDI of 100, Megawatt per metre | Expected life loss using above equation |
| 0 | 35 | 63 | 31 (100%) |
| 1.6 | 35 | 63 | 31 (100%) |
| 5 | 23 | 32 | 10 (32%) |
| 8 | 20 | 23 | 4 (13%) |
| 10 | 17.5 | 18 | 1 (3%) |
| 15 | 12 | 8.6 | 0 (0%) |
Estimating life loss by bushfire area
There are two satisfactory sources of information about the relationship between annual fuel reduction burning and subsequent bushfire area loss. The first is a modelling study for heath lands in SW Tasmania and the second is the actual result of 45 years of fuel reduction in the SW of Western Australia. The two studies support each other nicely and it is as reassuring to have a modelling result validated by actual experience, as it is to have a theoretical underpinning for actual experience. The assumption is made that the losses are proportional to the area burnt.
Table 3 predicts the lives lost using Black Saturday as the benchmark using the modelled data from Tasmania.
| Prescribed burning program | 0% pa | 5% pa | 8% pa | 10% pa | 15% pa |
| Predicted bushfire loss | 100% | 66% | 45% | 33% | 20% |
| Expected life loss for the Black Saturday event | 173 | 115 | 80 | 59 | 35 |
Table 4. Predicts the life loss for a Black Saturday event but using the actual data from Western Australia.
| Prescribed burning program | 0% pa | 5% pa | 8% pa | 10% pa | 15% pa |
| Predicted bushfire loss | 100% | 45% | 27% | 19% | 5% |
| Expected life loss for the Black Saturday event | 173 | 78 | 48 | 33 | 9 |
The lower expected life loss for the actual West Australian data could be due to the effectiveness of fuel reduction at decreasing the fire intensity as well as the area burnt. In fact if the decrease in fire intensity is calculated for the resulting fuel levels then the actual and Tasmanian modelled expectations are much closer.
Discussion
This note makes many assumptions which in future analysis should be validated or amended. The assumptions used in this analysis are that essentially all fires burn through the same populations and the relationship between Byram fire intensity and life loss is simple and thus the data for the life loss-intensity relationship should cover all fires and intensities.
Irrespective of the weakness of these assumptions this note gives an indication of the expected efficacy of fuel management by burning in protecting our communities against bushfire life loss.
It is safe to conclude that it is not until we undertake a 15% pa burning program that we could expect no life loss. The gain at 5%pa is small but by 8% pa it is becoming very much better. There is informed opinion amongst fire managers that 10 – 12% is the optimum for fire protection. This analysis supports their hard won wisdom.